The development of suitable cyclic loading and boundary conditions for ballast box tests

Abstract

Laboratory tests on ballast give insight into the behaviour and performance of the ballast layer under passenger and heavy haul traffic. It is important, however, to ensure that the simulation of train loads on the ballast layer in the laboratory represents in situ loading conditions. With adequate, representative loading in the laboratory, the settlement, stiffness, modulus and overall performance of the ballast layer can be estimated and predicted in the future. However, a review of current laboratory tests on ballast reveals that these do not employ the approximate in situ loading conditions. Furthermore, adequate ballast confinement in laboratory tests should model the confinement along the track as this gives an indication of an ideal response of the ballast layer in situ as well as its impact on track structure deterioration. The objective of this study was therefore to develop suitable cyclic loading and boundary conditions for ballast box tests in the laboratory which would represent similar conditions in the field. Literature studies reveal the typical train loading pattern at the rail seat (referred to as the Field loading (FL) pattern) which comprises of four load pulses with frequency depending on the wheel configuration. The FL pattern was compared with four alternative haversine loading patterns namely Laboratory Loading (Lab. L), Impulse Haversine Loading (IHL), Haversine Loading (HL) and Adjusted Haversine Loading (AHL) patterns. As a result of the complex shape of the FL pattern, a suitable alternative loading pattern was determined by comparing the rates of axial deformation caused by the FL pattern and an alternative loading pattern. It was found that the AHL pattern caused a similar rate of strain accumulation as the FL pattern. Furthermore, increasing the rest period interval between load cycles decreases the rate of ballast settlement. Suitable boundary conditions for the ballast layer were assessed by varying the level of lateral confinement while monitoring the rate of strain accumulation and the degree of ballast breakage. A fully confined ballast layer (100 % lateral confinement) produced limited axial deformation and less ballast breakage in comparison to a ballast layer with no lateral confinement. Ballast settlement increased by 150 % when the lateral confinement in the ballast box tests were reduced from 100 % to 0%. The changes in vertical pressure at the base of the ballast layer were investigated at different levels of confinement. Average vertical pressures of ~4800 kPa was observed for 0 % laterally confined ballast, while average vertical pressures of ~3800 kPa was observed for a fully confined ballast layer. The laboratory loading pattern developed in this research could provide accurate predictions of the long term behaviour of ballast as well as aiding the planning for subsequent ballast maintenance interventions based on realistic and accurate laboratory test results.